The present invention relates to systems and methods for visualizing subsurface regions of samples, and more specifically, to a frequency domain optical coherence reflectometer and frequency domain optical coherence tomography (OCT) device that provides internal depth profiles and depth resolved images of samples.
Optical coherence reflectometry/tomography involves splitting an optical radiation into at least two portions, and directing one portion of the optical radiation toward a subject of investigation. The subject of investigation will be further referred to as a “sample”, whereas the portion of optical radiation directed toward the sample will be further referred to as a “sample portion” of optical radiation. The sample portion of optical radiation is directed toward the sample by means of a delivering device, such as an optical probe. Another portion of the optical radiation, which will be further referred to as “reference portion”, is used to provide heterodyne detection of the low intensity radiation, reflected or backscattered from the sample.
Typically, any optical coherence reflectometer or OCT device is specified by a longitudinal (in-depth) range of interest, whereas the longitudinal range of interest and the sample overlap, at least partially. The longitudinal range of interest includes a proximal boundary and a distal boundary, and in time domain systems is equivalent to the longitudinal scanning range. In traditional time domain optical coherence reflectometry, at every moment only a small part of the sample portion of the optical radiation, reflected or backscattered from some point located inside the boundaries of the longitudinal range of interest is utilized. In-depth profiling of the sample is provided by introducing a variable optical path length difference for the sample and reference portions of the optical radiation.
A well known version of time domain optical coherence reflectometry and tomography is the “common path” version, also known as autocorrelator or Fizeau interferometer based OCR/OCT. In this version, the reference and sample portions of the optical radiation do not travel along separate optical paths. Instead, a reference reflection is created in the sample optical path by introducing an optical inhomogenuity in the distal part of the delivering device, the inhomogenuity serving as a reference reflector. Resulting from that, the reference and sample portions of the optical radiation experience an axial shift only. The distance between the reference reflector and the front boundary of the longitudinal range of interest will be considered here as “reference offset”. The entire combination of the sample portion of the optical radiation and axially shifted reference portion is combined with the replica of the same combination, shifted axially, so the reference portion of one replica has a time of flight (or optical path length) matching that of the sample portion of another replica. These portions interfere in a very similar way to the traditional “separate path” time domain optical coherence reflectometry/tomography embodiments. The interference signal is formed by a secondary interferometer, the two arms of which have an optical length difference (“interferometer offset”) equal to the reference offset. By scanning an optical delay between the two replicas, a time profile of the interference signal is obtained, which represents the in-depth profile of the coherent part of the reflected sample optical radiation. The later is substantially equivalent to the profile obtained in traditional separate path embodiments.
Common path frequency domain reflectometry/tomography has a lot of intrinsic advantages over separate path frequency domain reflectometry/tomography. These advantages are based on the fact that reference and sample portions of the optical radiation propagate in the same optical path and therefore experience substantially identical delay, polarization distortions, optical dispersion broadening, and the like. Therefore, the interference fringes are insensitive to the majority of the probe properties, including the optical fiber probe length, dispersion properties and polarization mismatch. In separate path frequency domain reflectometry/tomography, the length and dispersion of the sampling arm should be closely matched with the reference arm and the polarization mismatch should be prevented (using PM fiber or other means) or compensated (using polarization diversity receiver or other means).
The optical spectrum of the combined reference and sample portions of the optical radiation, both in the separate path and the common path reflectometry and OCT designs has all necessary information about the in-depth coherent reflection profile by including a component that is Fourier conjugate of the in-depth profile of the sample. Thus, the profile is capable of being extracted from Fourier transformation of the optical spectrum of the combined optical radiation.
Fourier transformation of the optical spectrum of the reference and sample optical radiation combination is actually well known and has been utilized in frequency domain optical coherence reflectometry and tomography (also known as spectral domain and Fourier domain) since 1995. In frequency domain optical coherence reflectometry, the reference and sample portions of the optical radiation have a substantially similar optical path. The optical spectrum of the combined optical radiation can be registered using parallel means (such as a spectrograph) or sequential scanning means using a swept frequency optical source.
However, it took several years for the scientific community to realize that frequency domain optical coherence reflectometry/tomography has a fundamental, major advantage in signal-to-noise ratio (SNR) over traditional time-domain reflectometry/tomography. The frequency domain reflectometry/tomography SNR advantage can be explained by a simultaneous use of light coming back from all in-depth pixels, whereas in time domain reflectometry/tomography, only light from one in-depth pixel is used at a time and all the rest is wasted. Therefore, the SNR for frequency domain reflectometry/tomography is capable of being improved by a factor equal to the number of in-depth pixels (which for a system with moderate in-depth resolution of 15 μm and scanning depth of 2 mm will be a factor of 133). It should be also noted that this advantage increases with improving in-depth resolution for the same depth, reaching a factor of 1000 for 2 μm resolution.
Common path frequency domain optical coherence reflectometry and tomography are well known in the art. However, previously known devices typically employ an optical layout where reference reflection occurs in the vicinity of the sample. In these devices the combination of reference and sample reflection is directly spectrally analyzed without any additional optical processing, such as using an additional interferometer. This approach works very well if stable reference reflection can be obtained from a point axially close to the sample. Unfortunately, in many situations, and in particular, in a probe design for medical application it is very difficult or even impossible to obtain reference reflection from the vicinity of the sample and instead, reference reflection can only be obtain from a point located far from the sample.
A limitation to such common path frequency domain OCR/OCT systems without a secondary interferometer is the great value of required spectral resolution of the frequency domain OCR/OCT processing engine. This limitation becomes especially important in medical applications. The problem is that even for miniature optical fiber endoscopic probes known in the art that use the optical fiber tip of the optical fiber probe as a reference element, the reference offset could be as big as 10 mm, since the optical fiber probe inevitably includes a lens system in its distal part. This distance may be greater if a bigger probe with a larger field of view is required, such as for laparoscopy. It is known that the larger the in-depth distance is between the most remote points involved in the optical interference (which is the reference offset plus intended depth range), the finer the spectral resolution of the system should be, in order to resolve the highest frequency spectral fringes.
The later can be illustrated referring to the spectrum of two pairs of pulses with different time separation. Each pair of pulses (for OCR/OCT corresponding to a pair of reflecting surfaces separated in depth) produces interference fringes in the spectrum. The frequency of spectral fringes increases accordingly with increasing of the delay between pulses. To restore the in-depth profile, the spectral resolution of the frequency domain OCR/OCT engine should be sufficient to resolve the most frequent fringes in the optical spectrum. In spatial-temporal terminology, the effective coherence length should be sufficient to provide interference between the most distant points. Therefore, a large reference offset creates unnecessary high spectral resolution requirements for the spectrometer or unnecessary strict instantaneous line width requirements for the tunable source. It also puts an additional burden on the data acquisition and real time signal processing system, where a several times increase of data flow is required for the same image acquisition rate. Additionally, the system design would require substantial changes if another probe with different reference offset is needed. All of the described is capable of making questionable the advantage of using common path topology in a frequency domain OCR/OCT system.
One solution would be to add an additional interferometer in the manner known for time domain OCT/OCR systems. Unfortunately, applying frequency domain registration to earlier separate path OCR/OCT systems creates a serious problem—the “depth ambiguity problem” (also referred to as mirror artifact or depth degeneracy). The problem is well known and is associated with Fourier transformation's inability to differentiate between positive and negative depth coordinates in a case of the optical path difference for the interfering reference and sample portions of the optical radiation being reduced to zero. The same problem would arise for a common path frequency domain OCR/OCT system utilizing a secondary interferometer since in a system of this type, as discussed above, the interference signal is formed by reducing to zero the optical path difference for the interfering reference and sample portions of the two replicas of the optical radiation. There are several ways known to deal with the depth degeneracy problem, all of them being cost consuming and rather complicated for being used in a medical device.
Another limitation to previously known common path frequency domain reflectometry/tomography devices is that the registered interference signal is responsive only to the non-depolarized portion, or in other words, responsive only to the parallel-polarized component of the optical radiation reflected or backscattered from the associated sample. The portion of the optical radiation depolarized by the associated sample and reflected or backscattered from it (the cross-polarized component), does not produce interference fringes and is not registered. However, in many cases OCR/OCT images created from the depolarized portion of the optical radiation demonstrate enhanced contrast and could be successfully used for biomedical diagnostics.
As will be appreciated by those skilled in the art, the concept of “parallel-polarized” and “cross-polarized” is applied here for elliptical polarization. “Parallel-polarized” is used for components with elliptical polarizations having the same eccentricity, same orientation of the long axis (ellipse tilt angle), and same rotation direction for the electric field. “Cross-polarized” is used for components with elliptical polarizations having the same eccentricity, orthogonal orientation of the long axis, and opposite rotation direction for the electric field. As in the case of linear or circular polarization these parallel-polarized components produce strongest interference, while cross-polarized components do not interfere at all.
Thus, there exists a need for common path frequency domain OCR/OCT devices that use the advantages of a common path optical interferometer design together with the advantages of frequency domain registration of the optical spectrum of the combined reference and sample portions of the optical radiation, overcoming at the same time limitations of both approaches.
There also exists a need for common path frequency domain OCR/OCT devices that provide registration of the portion of the optical radiation depolarized by an associated sample, i.e. of the cross-polarized component of the optical radiation reflected or backscattered from an associated sample.